Comment by kasbah

To me the interesting thing is the logic programming "rules" and their overlap with game rules. Inspired by work at Stanford on the logic programming based Game Description Language [1] I implemented Tic Tac Toe in Datascript yesterday: https://github.com/kasbah/datascript-games/blob/e06a37025bf9...

I am still not clear whether there isn't a more succinct rule definition than what I have there. In the Stanford paper you have rules like:

   (<= (column ?n ?x)
     (true (cell 1 ?n ?x))
     (true (cell 2 ?n ?x))
     (true (cell 3 ?n ?x)))

But in Datascript I have to do much more rigmarole around shuffling the data around:

  [(column ?n ?x) 
    [?current "ident" "current"] 
    [?coord0 "type" "coord"] 
    [?coord1 "type" "coord"] 
    [?coord2 "type" "coord"]
    [?coord0 "m" 0] 
    [?coord1 "m" 1] 
    [?coord2 "m" 2] 
    [?coord0 "n" ?n] 
    [?coord1 "n" ?n] 
    [?coord2 "n" ?n] 
    [?coord0 "name" ?key0] [?current ?key0 ?x]
    [?coord1 "name" ?key1] [?current ?key1 ?x]
    [?coord2 "name" ?key2] [?current ?key2 ?x]]

I don't know if this is down to Datascript/Datomic Datalog limitations or more the limitations of my understanding.

How do you approach your experiments? If you have any of your work to share or some tips on what I am doing I'd be very interested.

[1]: https://www.cs.uic.edu/~hinrichs/papers/love2006general.pdf